We have just published a paper in Cities (you can download if for free here) that looks at the history of policies around tenure mix in New Zealand, and draws on de-identified integrated government data to understand how public housing is distributed across New Zealand, and how the proportion of public housing in a community affects the health of tenants living there.
This is an important thing to look at because it is commonly assumed (including in NZ, as we found in recent research) that living in an area with a lot of other public housing is bad for tenants. This belief can help justify building less public housing on a site – at a time when we do not have homes for almost 27,000 households who are eligible for public housing. For example, some housing providers argue that the proportion of public housing in a new development should be limited to 20-30%.
The distribution of public housing
We found that, in 2013, most public housing tenants live in communities with relatively low proportions of public housing tenants in the population – both at the meshblock level (housing a median of 78 people; i.e., a street or building) and the census area unit level (housing a median of 1,863 people; i.e., one or two census area units make up a typical suburb). In the paper, we outline the different policies which, over the decades, have contributed to the distribution of public housing.
As the graph above shows, approximately 70% of public housing tenants lived in a meshblock where 50% or less of the population lived in public housing, and 11% percent lived in a meshblock where 80% or more of the population lived in public housing. About 85% of public housing tenants live in a census are a unit where less than 50% of the population lived in public housing. 0.1% lived in a census area unit with more than 80% public housing.
How the proportion of public housing in a community affects tenants’ health
We looked at the relationship between the percentage of public housing tenants in meshblocks, or census area units, where public housing tenants live, and their hospitalisations five years later. The five-year gap enables us to monitor long-term effects and takes into account that, if the proportion of public housing does affect health, this is likely to take some time to show up in the data.
In the graph below, the different levels of public housing are grouped together in one-percent-point groups. For example, the leftmost point in the graph below shows the mean number of hospitalisations for all public housing tenants living in meshblocks where 0-1% of the population were public housing tenants. The size of each point reflects the sample size living in areas with that proportion of public housing tenants.
The graph above shows that as the proportion of public housing tenants in the population increases, the hospitalisation rate generally decreases slightly, with our model results implying there is a minimum at 49%. Those living in areas with less than 15% public housing tenants have 20% more hospitalisations on average than those living in areas with 15% to 70% public housing tenants. The highest rates of hospitalisation occur in people living in meshblocks with very low, or very high fractions of public housing tenants. While statistically significant, the effect size is modest, with our model predicting that those living in meshblocks with 49% public housing tenants would have on average 0.05 fewer hospitalisations than those living in meshblocks with 10% public housing tenants.
We conducted the same analysis at the census area unit level, and found broadly the same pattern. As explained in greater depth in the paper, we found similar results for mental health outpatient visits and pharmaceutical dispensing.
Our research leads us to conclude:
Contrary to common assumptions, the proportion of public housing in meshblocks and census area units had only a minimal effect on the health outcomes of the public tenants who lived there five years later. Living in a community where more public housing tenants were resident had a minor, but positive impact on health outcomes for public housing tenants. Planning for larger proportions of public housing in streets and neighbourhoods would therefore not only be beneficial for public housing tenants, it would result in more public housing overall, in a time where public housing is in great demand. The issue of public housing concentration and residential mix has received considerable attention in NZ’s history, especially in recent years; we would argue that the question of providing adequate housing, incomes and health and other social services to enable public housing tenants and other low-income people to live a life of dignity is of much greater importance.Chisholm, E., Robertson, O., Howden-Chapman, P., & Pierse, N. (2022). Does the proportion of public housing tenants in a community affect their wellbeing ? Results from New Zealand: A retrospective cohort study using linked administrative data. Cities, 131(103916), 1–10. https://doi.org/10.1016/j.cities.2022.103916